Related Experiment Video
Updated: Jun 27, 2026
05:58
Retinal Detachment Model in Rodents by Subretinal Injection of Sodium Hyaluronate
Published on: September 11, 2013
20.8K
A retinal detachment based strabismus detection through FEDCNN
Ayesha Jabbar1, Muhammad Kashif Jabbar1, Tariq Mahmood2,3
1College of Electronics and Information Engineering, Shenzhen University, Shenzhen, 518060, China.
Scientific Reports
|October 6, 2024
Summary
This study introduces a novel approach using Convolutional Neural Networks (CNNs) and eye-tracking data to accurately diagnose ocular strabismus. The developed FedCNN model achieved 95.2% accuracy, improving diagnostic precision for eye conditions.
Area of Science:
- Ophthalmology
- Computer Science
- Medical Diagnostics
Background:
- Ocular strabismus is a significant risk factor for amblyopia and vision loss.
- Current diagnostic methods for strabismus lack sufficient accuracy and reliability, despite advancements in eye-tracking technology.
Purpose of the Study:
- To enhance the accuracy and reliability of strabismus diagnosis.
- To develop an automatic strabismus detection system integrating novel algorithms.
Main Methods:
- A novel FedCNN model was proposed, combining Convolutional Neural Networks (CNNs) with eXtreme Gradient Boosting (XGBoost).
- The model utilizes Gaze deviation (GaDe) images to capture dynamic eye movements for precise feature extraction.
- Eye-tracking datasets from subjects were employed for training and validation.
Main Results:
- The FedCNN model achieved a diagnostic accuracy of 95.2% for strabismus detection.
- The CNN's detailed connection layer effectively selected features crucial for strabismus recognition.
- The proposed method demonstrated high precision in diagnosing strabismus.
Conclusions:
- The developed FedCNN model significantly improves the accuracy of strabismus diagnosis.
- This approach has the potential to transform eye disease diagnostics for a substantial number of patients.
- The integration of CNNs and eye-tracking data offers a promising direction for automated ophthalmological diagnostics.
